Worrying about the future: towards evidence-based prognosis in anxiety disorders.

Anxiety disorders are a group of psychiatric disorders with high impact on patients and society due to its high prevalence, low treatment usage, high levels of comorbidity and disability. The clinical course is often chronic. In routine care, most emphasis lies on immediate treatment effects, whereas a focus on prognosis would be much appreciated. The first aim in this thesis was to expand on the knowledge about factors that impact the clinical course in anxiety disorders. To describe these factors an ordering across susceptibility markers, diagnostic markers, prognostic markers and predictive markers can be used. The findings in this thesis underscore the previous findings that low socio-economic status, female gender and younger age are related to subsequent development of anxiety disorders and are thereby a susceptibility marker. Lifetime history of anxiety disorders and family history of psychiatric disorders were used in chapter 5 to assign subjects to at-risk stages. These risk factors can be regarded as susceptibility markers for future development of anxiety disorders. By analysing and integrating the available literature on the topic of treatment resistance in anxiety disorders a new consensus definition was created for treatment resistant anxiety disorders. This definition included at least one first-line pharmacological treatment and one first-line psychotherapeutic treatment, provided for an adequate duration (at least 8 weeks) and according to local protocol. Additional criteria for lack of response and minimal symptom severity were formulated. This new definition could be used as a diagnostic marker for the more advanced stages of anxiety disorders. Duration of anxiety disorders could be seen as a diagnostic marker for the different stages of anxiety disorders, with longer duration indicating a more advanced stage of anxiety. The presence of subthreshold anxiety symptoms along with the presence of susceptibility markers seems an adequate diagnostic marker for the prodromal stages of anxiety disorders. Presence of chronic somatic diseases was found to be a poor prognostic marker in anxiety disorders and depressive disorders. In all stages of anxiety disorders, presence of psychiatric comorbidity was clearly associated with markedly worse outcomes at follow-up. Presence of psychiatric comorbidity should be seen as a poor prognostic marker. A measurement tool for the degree of treatment resistance in anxiety disorders was found to be moderately associated with the subsequent course during treatment. This treatment resistance score can be regarded as a predictive marker for anxiety disorders. The second main aim for this thesis was to improve predictions for the clinical course of anxiety disorders by assessing predictive properties of different prediction models. The first model presented is the clinical staging model (chapter 5). This prognostic model is based on a clinician-opinion model for disease progression in anxiety disorders. The second model presented (chapter 6) is a predictive model in which a dimensional measurement instrument was developed based on literature-derived criteria for advanced progression in anxiety disorders. The final prognostic model presented is a data-driven model (chapter 7) for naturalistic course in anxiety disorders. Overall, the accuracy of the predictions in the models presented were consistently moderate. Comparison with earlier prediction models in anxiety disorders are difficult as previous attempts are scarce. First, the results and integration of this thesis warrant applying the staging paradigm in anxiety disorders. Routinel care should therefore at least consist of a combined clinical assessment of previous duration, anxiety severity, presence of disability and presence and severity of comorbidity. Second, when assessing anxiety disorder patients, previous treatments and its effects should be assessed. The definition for treatment resistant anxiety disorders provided in this thesis could be used to align diagnostic criteria in this group and provide alignment with treatment guidelines.